ciedesigns
Español Discuss your project
  • Public beta
  • Accessibility technology
  • WebMCP exploration

EasyACR

Designing and building an AI agent-assisted accessibility tool assisting in gathering and recording accessibility reporting evidence.

Sector
Accessibility technology
Dates
August 2026–ongoing
Role
End-to-end product designer
Scope
Strategy, identity, product design, system, validation, front end
EasyACR screens showing the public site, scan workflow, issue review, and draft accessibility-evidence experience.

01 — Context

A focused public beta for accessibility evidence

EasyACR is intended for software teams preparing accessibility evidence for procurement and review. The public beta connects bounded website scanning, issue investigation, and a structured draft-evidence workflow. It does not present automated findings as certification or legal advice, but provides an ACR framework to be completed by a human reviewer.

02 — Challenge

Make complex evidence work understandable without overstating the result

Accessibility review combines automated findings with human judgment. The product needed to make that boundary unmistakable while organizing technical results into a workflow a team could understand, revisit, and prepare for further review — and expose useful WebMCP actions without letting an agent exceed the user-authorized workflow.

EasyACR dashboard listing recent public-site scans and automated findings that still require human review.
EasyACR public site section explaining what the product does and does not do.

03 — Approach

Design the product, system, and implementation together

The engagement covered research, product strategy, information architecture, visual identity, a reusable design system, prototypes, accessibility validation, and front-end implementation. The interface separates automated evidence from manual review and applies repeatable components across public, authenticated, report, and draft-evidence views.

  • Foundations

    Product features and safety were defined before the visual identity and responsive component system.

  • Components

    Both the web and MCP versions of scanning, findings, and draft-evidence flows are built from one component set.

  • Flows

    Some checks, including manual WCAG evaluation, must be completed by the user.

04 — Outcome

A live beta and a practical WebMCP learning platform

The public beta launched in early September 2026. It provides a working product experience for public-site scans and draft WCAG evidence while noting explicit review gates. This afforded us the opportunity to learn how WebMCP can support user-initiated actions inside a web product.